Some people when deploying Odoo do so by the simple route of running odoo-bin, which does not allow for having an infrastructure deployed to balance the load between several of these servers. One of the problems is the storage of the user's session data, which by default is written into the Odoo data folder, invalidating the possibility of load balancing between 2 (or more) because the sessions would not be found between them.

I created this repository to provide a Dockerfile prepared with Odoo 13 with these characteristics:

• It uses uWSGI to run Odoo, with automatic scaling of the workers.
• It uses NGINX as a reverse proxy to connect to the uWSGI using its own protocol and through a UNIX socket.
• It provides the module statics directly from NGINX without going through Odoo.
• NGINX has support for brotli and gzip, considerably reducing the load wait for clients.
• Use a module that add Redis support to store HTTP Sessions.
• Use a module that allow save the attachments to PostgreSQL using Large Object.
• A Monkey Patch is applied to Odoo to be executed using Gevents, which increases its performance with less resource consumption.

Details to be taken into account:

• When Odoo runs with uWSGI it doesn't raise the LongPolling and Cron service, the image is prepared to be run separately (at least 1) with the environment variable ODOO_MODE=longpoll-cron which tells it to run only the LongPolling and Cron.
• To install new modules it is advisable to use the environment variables in a new image ODOO_MODE=install-update-modules together with MODULES_TO_INSTALL=modulo1,modulo2,modulo3....

Simple, use the docker build command.

The entrypoint of this image use some enviroment variables:

The image start mode can be:

• longpoll-cron: Init the Long Polling + Cron of Odoo.
• install-update-modules: Install the modules defined in MODULES_TO_INSTALL
• uwsgi-server: Init uWSGI + Nginx to serve

If you want add custom modules, Odoo is configured to search on /usr/src/app/extra-modules, you can mount a volume inside it.

• https://github.com/odexperts/odoo-cloud
• https://github.com/camptocamp/odoo-cloud-platform